Neurobehavioral correlates of obesity are

largely heritable
Uku Vainika,b, Travis E. Bakera,c, Mahsa Dadara, Yashar Zeighamia, Andréanne Michauda, Yu Zhanga,
José C. García Alanisa,d, Bratislav Misica, D. Louis Collinsa, and Alain Daghera,1
a
Montreal Neurological Institute, McGill University, Montreal, QC H3A 2B4, Canada; bInstitute of Psychology, University of Tartu, Näituse 2, 50409 Tartu,
Estonia; cCenter for Molecular and Behavioral Neuroscience, Rutgers University, Newark, NJ 07102; and dNeuropsychology Section, Experimental and
Biological Psychology, Department of Psychology, Philipps University of Marburg, 35032 Marburg, Germany

Edited by Daniel H. Geschwind, University of California, Los Angeles, CA, and accepted by Editorial Board Member Michael S. Gazzaniga August 6, 2018
(received for review October 20, 2017)

Recent molecular genetic studies have shown that the majority of reports of cortical thickness patterns associated with obesity have
genes associated with obesity are expressed in the central nervous been inconsistent (12, 13). As a prerequisite to our goal of
system. Obesity has also been associated with neurobehavioral factors ascertaining the heritability of brain-based vulnerability to obesity,
such as brain morphology, cognitive performance, and personality. we sought to extend previous neurobehavioral findings in a large
Here, we tested whether these neurobehavioral factors were associ- multifactor dataset from the Human Connectome Project (HCP).
ated with the heritable variance in obesity measured by body mass We also measured volumetric estimates of medial temporal lobe
index (BMI) in the Human Connectome Project (n = 895 siblings). Phe- and subcortical structures, which have been implicated in appeti-
notypically, cortical thickness findings supported the “right brain hy- tive control (e.g., ref. 14).
pothesis” for obesity. Namely, increased BMI is associated with decreased The main goal was to assess whether the aforementioned
cortical thickness in right frontal lobe and increased thickness in the left obesity–neurobehavioral associations are of genetic or environ-

NEUROSCIENCE
frontal lobe, notably in lateral prefrontal cortex. In addition, lower thick- mental origin. Recent evidence from behavioral and molecular
ness and volume in entorhinal-parahippocampal structures and increased genetics suggests that there is considerable genetic overlap
thickness in parietal-occipital structures in participants with higher BMI among obesity, cognitive test scores, and brain imaging findings
supported the role of visuospatial function in obesity. Brain morphome- (15–20). However, the evidence so far is not comprehensive
try results were supported by cognitive tests, which outlined a negative across all neurobehavioral factors discussed. A recent paper
association between BMI and visuospatial function, verbal episodic mem- assessed the heritability of obesity-associated regional brain
ory, impulsivity, and cognitive flexibility. Personality–BMI correlations volumes (21). However, the study did not analyze the heritability
were inconsistent. We then aggregated the effects for each neurobeha- of the association between brain and obesity. The latter analysis
vioral factor for a behavioral genetics analysis and estimated each factor’s is crucial for understanding whether brain anatomy and obesity
genetic overlap with BMI. Cognitive test scores and brain morphometry
could have a genetic overlap, which would suggest that the
had 0.25–0.45 genetic correlations with BMI, and the phenotypic correla-
heritability of vulnerability to obesity is expressed in the brain.
tions with BMI were 77–89% explained by genetic factors. Neurobeha-
In addition, we sought to estimate the genetic overlap between
vioral factors also had some genetic overlap with each other. In
the different BMI-related neurobehavioral factors. Performance
summary, obesity as measured by BMI has considerable genetic overlap
with brain and cognitive measures. This supports the theory that obesity
is inherited via brain function and may inform intervention strategies. Significance

brain morphology | cortical thickness | cognition | twins | body mass index Obesity is a widespread heritable health condition. Evidence
from psychology, cognitive neuroscience, and genetics has
proposed links between obesity and the brain. The current
O besity is a widespread condition leading to increased mor-
tality (1) and economic costs (2). Twin and family studies
have shown that individual differences in obesity are largely
study tested whether the heritable variance in body mass in-
dex (BMI) is explained by brain and behavioral factors in a large
brain imaging cohort that included multiple related individuals.
explained by genetic variance (3). Gene enrichment patterns
We found that the heritable variance in BMI had genetic cor-
suggest that obesity-related genes are preferentially expressed in relations 0.25–0.45 with cognitive tests, cortical thickness, and
the brain (4). While it is unclear how these brain-expressed genes regional brain volume. In particular, BMI was associated with
lead to obesity, several lines of research show that neural, cog- frontal lobe asymmetry and differences in temporal-parietal
nitive, and personality differences have a role in vulnerability to perceptual systems. Further, we found genetic overlap be-
obesity (5, 6). Here, we seek to test whether these neuro- tween certain brain and behavioral factors. In summary, the
behavioral factors could explain the genetic variance in obesity. genetic vulnerability to BMI is expressed in the brain. This may
In the personality literature, obesity is most often negatively inform intervention strategies.
associated with conscientiousness (self-discipline and orderli-
ness) and positively with neuroticism (a tendency toward nega- Author contributions: U.V., T.E.B., B.M., and A.D. designed research; U.V., T.E.B., M.D.,
tive affect) (7). In the cognitive domain, tests capturing executive Y. Zeighami, A.M., Y. Zhang, J.C.G.A., D.L.C., and A.D. performed research; U.V., T.E.B.,
M.D., Y. Zeighami, A.M., Y. Zhang, J.C.G.A., and D.L.C. analyzed data; and U.V., T.E.B.,
function, inhibition, and attentional control have a negative as- A.M., and A.D. wrote the paper.
sociation with obesity (5–8). Neuroanatomically, obesity seems The authors declare no conflict of interest.
to have a negative association with the gray matter volume of
This article is a PNAS Direct Submission. D.H.G. is a guest editor invited by the
prefrontal cortex and, to a lesser extent, the volume of parietal Editorial Board.
and temporal lobes, as measured by voxel-based morphometry This open access article is distributed under Creative Commons Attribution-NonCommercial-
(9). It has also been suggested that structural and functional NoDerivatives License 4.0 (CC BY-NC-ND).
asymmetry of the prefrontal cortex might underlie overeating 1
To whom correspondence should be addressed. Email: alain.dagher@mcgill.ca.
and obesity (10). For genetic analysis, cortical thickness estimates This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
of brain structure from magnetic resonance imaging (MRI) have 1073/pnas.1718206115/-/DCSupplemental.
been preferred over volumetric measures (11). However, to date,

www.pnas.org/cgi/doi/10.1073/pnas.1718206115 PNAS Latest Articles | 1 of 6

on cognitive tests and personality must originate from the brain
(e.g., ref. 22) and, therefore, personality and cognition could be
expected to explain brain–morphometry associations with BMI
(6). However, brain–behavior associations are far from certain
(23), and even different measurement traditions in both behavior
(personality and cognitive tests) and brain morphometry (cortical
thickness or brain volume) are often conceptualized as providing
independent sources of information (7, 11). Documenting the
degree of genetic overlap between behavioral and brain measures
would shed light on whether similar underlying processes lead to
obesity’s associations with different neurobehavioral factors.
Taken together, the goal of the current analysis was to use a large
multifactor dataset to analyze the heritability of the associations
between obesity and brain/behavior. We further tested genetic
overlap between the different neurobehavioral factors themselves.

Results
Background. We analyzed data from 895 participants from the
Human Connectome Project S900 release (24), including 111 pairs
of monozygotic twins and 188 pairs of dizygotic twins and siblings.
Similar to many previous reports (3) we modeled BMI heritability
with the AE model (A, additive genetics; E, unique environment),
as opposed to the ACE model (C, common environment), as AE
had the lowest Akaike Information Criterion (Dataset S1, section
9). BMI heritability was A = 71% [95% CI: 61%; 78%], which is
close to the published meta-analytic estimate (A = 75%, ref. 3).
In all analyses below, we controlled for age, gender, race,
ethnicity, handedness, and evidence of drug consumption on day
of testing, which mostly associated with BMI (SI Appendix, SI
Results and Fig. S2). When presenting and interpreting pheno-
typic associations, we controlled for family structure to avoid
inflated effect sizes and SEs (e.g., ref. 25). The behavioral ge-
netics analysis did not control for family structure, since this
information is needed for modeling heritability. As socio-economic
status (SES) is intertwined with cognitive test scores (26), person-
ality (27), and brain morphometry (28), we also present phenotypic
associations controlling for SES (education and income) in SI Ap-
pendix, Supplementary Material. All in-text P values are provided
without correcting for multiple comparisons. False discovery rate
(FDR) correction was applied when screening for features within
cognitive, personality, and brain factors (Figs. 1 and 2).

Cognitive and Personality Factors. BMI was negatively correlated

with the following tests of executive function: cognitive flexibility,
fluid intelligence, inability to delay gratification, reading abilities,
and working memory. Intriguingly, the strongest effects were
present for nonexecutive tasks measuring visuospatial ability and
verbal memory (Fig. 1A). These tasks remained associated with
BMI after controlling for SES; controlling for SES reduced the
number of executive function tests involved with BMI to cogni-
tive flexibility and inability to delay gratification (SI Appendix,
Fig. S3A, Left). No personality test score correlated with BMI
when FDR correction was applied (Fig. 1B).

Brain Morphology. Cortical thickness was estimated from each T1-

weighted MRI using CIVET 2.0 software (29). Parcel-based analysis
identified negative associations with BMI in right inferior lateral
frontal cortex and bilateral entorhinal-parahippocampal cortex
(Figs. 2A and 3A). Positive associations with BMI were found with
the left superior frontal cortex, left inferior lateral frontal cortex,
and bilateral parietal cortex parcels. Controlling for SES did not
change these results (SI Appendix, Fig. S4A, Left). The frontal lobe
Fig. 1. Associations between BMI and cognitive test scores (A) and personality
asymmetry in the BMI association (thinner on the right, thicker on
traits (B). Error bars represent 95% confidence intervals. See Dataset S1, sec-
the left) mostly involved the inferior lateral prefrontal areas, such as tion 1 for explanation of cognitive tests. Numerical values are reported in
inferior frontal gyrus. Dataset S1, section 2. EF, executive function; FDR, false discovery rate; FFM,
Regional brain volumes were measured for estimation of brain Five-Factor Model; Imp, (lack of) impulsivity; Lang, language; Mem, memory;
morphology–obesity associations in brain structures not covered Neg, negative affect; Perc, perception; PWB, psychological well-being; Soc,
by the CIVET cortical thickness algorithm. Medial temporal social relationships; SSE, stress and self efficacy; WM, working memory.

2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1718206115 Vainik et al.

 association between BMI and lower volume of the entorhinal
cortex bilaterally and a positive association of left amygdala
volume with BMI (Figs. 2B and 3B). No subcortical region had a
significant association with BMI, and results did not change
when controlling for SES (SI Appendix, Fig. S4B, Left).

Creating Poly-Phenotype Scores. We performed dimension re-

duction for heritability analyses to reduce measurement noise
and avoid multiple testing with redundant measures. Similar to
other recent papers (20, 27), we used the weights of each indi-
vidual feature within a neurobehavioral factor (personality test,
cognitive test, brain parcel) to create an aggregate BMI risk
score or poly-phenotype score (PPS). This is similar to the
polygenic score approach in genetics, where the small effects of
several polymorphisms are aggregated to yield a total effect
score (15, 19, 20, 27). We used the correlation values as weights
to multiply each participant’s scaled measurements and aggre-
gated the results into a single composite variable, the PPS. The
PPS reflects the total association of each neurobehavioral factor
with BMI. To avoid overfitting, we assigned each 10% of par-
ticipants the PPS weights obtained from the other 90% (see SI
Appendix, Data Analysis for details).
The associations between BMI and the PPS-s for cognition

NEUROSCIENCE
(correlation with BMI: r = 0.16, P < 0.001, n = 798) and per-
sonality (r = 0.08, P = 0.017, n = 888) are slightly higher than the
meta-analytic estimates of the pooled association between BMI
and cognitive test scores (r = 0.10, ref. 8) and personality factors
(r = 0.05, ref. 8). BMI had stronger associations with the PPS-s

Fig. 2. Associations between BMI and brain morphometry. (A) Cortical thick-
ness. (B) Medial temporal and subcortical regional brain volume. Error bars rep-
resent 95% confidence intervals. Numerical values are reported in Dataset S1,
section 2. FDR, false discovery rate; Fro, frontal, Ins, insula; L, left; MTL, medial
temporal lobe; Occ, occipital; Par, parietal; R, right; SC, subcortical; Tem, temporal.

Fig. 3. Brain maps of the associations between BMI and cortical thickness
lobe and subcortical volumes were individually segmented and (A) and medial temporal and subcortical regional brain volume (B) on a
measured by registering each brain to a labeled atlas using standard brain template in Montreal Neurological Institute space. Values are
ANIMAL software (30). Volumetric results demonstrated an the same as in Fig. 2. Color bar applies to both subplots. L, left; R, right.

Vainik et al. PNAS Latest Articles | 3 of 6

for cortical thickness (r = 0.26, P < 0.001, n = 591) and medial Genetic Overlap Between Neurobehavioral Factors. Phenotypically,
temporal brain volume (r = 0.23, P < 0.001, n = 594). There certain PPS-s had small but significant intercorrelations (SI Ap-
was no association between BMI and subcortical brain volume pendix, Fig. S11, upper triangle). After FDR correction, we were
(r = −0.05, P = 0.169, n = 828). To test the generalizability of the able to find two genetic correlations between PPS-s of cognition
PPS approach, we used weights obtained from the full S900 re- and cortical thickness (rg = 0.35), as well as cognition and per-
lease (SI Appendix, Fig. S3, Right and SI Appendix, Fig. S4, Right) sonality (rg = 0.33, SI Appendix, Fig. S11, lower triangle). Taken
to test PPS–BMI correlation among the unseen additional par- together, while the neurobehavioral factors have mostly in-
ticipants in the S1200 release (referred to as S1200n, n = 236). dependent effects on BMI, cognitive test scores may have a small
Cortical thickness PPS had essentially unchanged effect size genetic overlap with brain structure and personality.
when correlated with BMI in S1200n (SI Appendix, SI Results and
Fig. S7). At the same time, cognitive and personality PPS-s were Discussion
less stable (SI Appendix, SI Results and Fig. S7), likely because Cortical thickness, medial temporal lobe volume, and cognitive
the smaller effect sizes of individual features need larger training measures all had covariation with BMI, and their effect on BMI
datasets to reduce inaccuracies, or that the true PPS-BMI effect was almost entirely heritable. Similarly, we found genetic cor-
size was too small to be found just within the S1200n sample. relations between obesity risk scores of cognition, cortical
thickness, and personality. Together, our results from a large
Heritability. Bivariate heritability was similarly conducted with the sample support the role of brain and psychological constructs in
AE model, since the main goal was to explain variance in BMI, explaining genetic variance in BMI.
for which AE was the best model. All PPS-s were found to be BMI correlated with increased cortical thickness in the left
highly heritable, with the A component explaining 36–79% of the prefrontal cortex and decreased thickness in the right prefrontal
variance (Fig. 4A and Dataset S1, section 10). Significant genetic cortex, supporting the “right brain” hypothesis for obesity (10). The
correlations (rg) were found between BMI and cognitive test effect was most prominent in the inferior frontal gyrus (Figs. 2A and
scores [rg = 0.25 (P = 0.002), cortical thickness (rg = 0.45, P < 3A). Only preliminary support for the right brain hypothesis has
0.001), and medial temporal brain volume (rg = 0.36, P < 0.001) been previously available (13). Right prefrontal cortex has been
(Fig. 4B and Dataset S1, section 11). The personality PPS genetic implicated in inhibitory control (22) and possibly bodily awareness
correlation with BMI was not significant (rg = 0.22, P = 0.052). (10). Many neuromodulation interventions (e.g., transcranial mag-
Molecular evidence relying on linkage disequilibrium score re- netic stimulation) aimed at increasing self-regulation capacity often
gression has reported effects of similar magnitude between target right prefrontal cortex. However, effects have also been
higher cognitive test scores and BMI (rg = −0.22, ref. 15; rg = −0.18, demonstrated in studies targeting left prefrontal cortex (31).
ref. 18). Environmental correlations (i.e., correlations between en- Cortical thickness results also highlighted the role of temporo-
vironmental variances) were small and not significant (Dataset S1, parietal perceptual structures in obesity. Namely, BMI was as-
section 11). As expected from high heritability of the traits and high sociated with bilaterally decreased thickness of the parahippocampal
genetic correlations, the phenotypic BMI–PPS correlations de- and entorhinal cortices, and with mostly right-lateralized increased
scribed in the previous sections were 77–89% explained by ge- thickness of parietal and occipital lobes. Volumetric results within
netic factors (Fig. 4C and Dataset S1, section 10). the medial temporal lobe supported the role of entorhinal cortex
The results broadly replicated when repeating the analysis with and also suggested that obesity is positively associated with the
just the top features within a PPS, suggesting that PPS-based volume of left amygdala. Emergence of the effects of the right pa-
findings summarize the effects of the underlying individual fea- rietal structures together with right prefrontal structures hint at the
tures (SI Appendix, Fig. S8). We further replicated the heritability role of the ventral frontoparietal network, thought to be especially
patterns in a separate analysis focused only on the additional important for detection of behaviorally relevant visual stimuli (32).
participants from the S1200 HCP release (SI Appendix, Fig. S9). The parahippocampal and entorhinal cortex are associated with
Additionally, controlling for SES (education and income) did not episodic memory and context mediation (33). Similarly, the hippo-
change the results for brain-based PPS-s. However, the estimates campus has been associated with the modulation of food cue re-
for cognitive test scores and personality became lower and not activity by homeostatic and contextual information, and hippocampal
significant in the S900 release (SI Appendix, Fig. S10). However, dysfunction is postulated to promote weight gain in the western diet
the same estimates were significant in the combined sample environment (34). The amygdala is implicated in emotional and ap-
S900+S1200, suggesting that the effects of cognition and person- petitive responses to sensory stimuli, including food cues (35).
ality were reduced but not eliminated when controlling for SES. Integrating these findings, one could envision a model where
obesity is associated with a certain cognitive profile (36). The
model starts with a hyperactive visual attention system attribut-
ing heightened salience to food stimuli, implicating the ventral
visual stream and amygdala. These signals are then less optimally
tied into relevant context by the parahippocampal and entorhinal
structures, and less well moderated (or filtered) by the prefrontal
executive system. This could result in consummatory behavior
driven by the presence of appetitive food signals, which are
ubiquitous in our obesogenic environment. An impaired re-
sponse inhibition and salience attribution model of obesity has
been suggested based on the functional neuroimaging literature.
Fig. 4. Heritability analysis of the association between PPS and BMI. (A) Namely, functional MRI studies have consistently identified obe-
Heritability of each trait. BMI has multiple estimates, since it was entered sity to associate with heightened salience response to food cues,
into a bivariate analysis with each PPS separately. (B) Genetic correlations coupled with reduced activation in prefrontal and executive sys-
between BMI and each PPS. The genetic correlations are positive, because
tems involved in self-regulation and top-down attentional control
the PPS-s are designed to positively predict BMI. (C) Heritability of the sig-
nificant phenotypic correlation between BMI and PPS. Horizontal lines de-
(e.g., ref. 35). A similar conclusion emerged from a recent resting
pict 95% confidence intervals. Cogn, PPS of cognitive tests; corr, correlation; state network analysis of the HCP data (37), in which obesity was
CT, PPS of cortical thickness; MTL, PPS of medial temporal lobe volume; Pers, associated with alterations in perceptual networks and decreased
PPS of personality tests; SC, PPS of subcortical structure volumes. activity of default mode and central executive networks.

4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1718206115 Vainik et al.

 This brain morphology-derived model has some support from 29. Young adults often experience “healthy or transitional obesity,”
cognitive tests. The role of prefrontal executive control is out- where clinical inflammation levels (45) and other cardiometabolic
lined by our finding of a negative association between BMI and comorbidities have not yet developed (46).
scores on several executive control tasks. Surprisingly, there was We found neurobehavioral PPS-s to have occasional phenotypic
no effect of motor inhibition as measured by the Flanker inhibitory and genetic correlations with each other. Here, it is hard to argue
task. A relation between obesity and reduced motor inhibition, while against pleiotropy playing a role. While one could reasonably ex-
often mentioned, has been inconsistent even across meta-analyses pect that at least part of the variation in cognitive performance
(7, 8). However, we found a relationship between decisional im- would be shaped by brain morphometry (22), it is also the case that
pulsivity, measured by delay discounting, and BMI, replicating pre- engaging in education leads to improvement in cognitive test scores
vious literature (6, 7, 18). While controlling for education reduced (26) and might also lead to changes in cortical thickness (47). The
the number of executive tasks associated with BMI, the overall small genetic overlap between cognition, cortical thickness, and
pattern remained the same, suggesting that education level is a proxy personality can probably be explained by common pleiotropic roots.
for certain executive function abilities. At the same time, integrating morphometry and cognitive findings
Intriguingly, BMI was found to be negatively associated with is difficult with this dataset.
spatial orientation and verbal episodic memory. These tasks tap From a practical point of view, our work suggests that evi-
into the key functions associated with entorhinal and para- dence from psychology and neuroscience can be used to design
hippocampal regions implicated in our study (33). Therefore, intervention strategies for people with higher genetic risk for
both cognitive and brain morphology features propose that the obesity. One way would be modifying neurobehavioral factors,
increased salience of food stimuli could be facilitated by dysre- e.g., with cognitive training, to improve people’s ability to resist
gulated context representation in obesity. the obesogenic environment (31, 36). Another path could be
Regarding personality, we were unable to find any questionnaire- changing the immediate environment to be less obesogenic (e.g.,
specific effects, notably with respect to neuroticism and conscien- ref. 48) so that individual differences in neurobehavioral factors would
tiousness, both often thought to be associated with obesity (5–7). be less likely to manifest. In any case, obesity interventions should not
There are potential explanations for this negative finding. First, the

NEUROSCIENCE
focus solely on energy content, but also acknowledge the certain
meta-analytical association between various personality tests and neurobehavioral profile that obesity is genetically intertwined with.
BMI is small (r = 0.05, ref. 7), for which we might have been un- The current analysis has limitations. Due to the cross-sectional
derpowered after P value correction. Second, controlling for family nature of the dataset, causality between neurobehavioral factors
structure likely further reduced the effect sizes (25). Third, the per- and BMI is only suggestive—longitudinal designs would enable
sonality–obesity associations tend to pertain to more specific facets better insight into the causal associations between brain morphol-
and nuances than broad personality traits (38), therefore, further ogy, psychological measures, and BMI or weight gain. BMI is a
analysis with more detailed and eating-specific personality measures crude proxy for actual eating behaviors or health status. In addition,
is needed in larger samples.
there were more normal-weight than obese participants. However,
All of the associations discussed here were largely due to
the 25% obesity rate in this sample is close to the published obesity
shared genetic variance between neurobehavioral factors and
rate of the state of Missouri (31.7%) and the United States (36.5%,
BMI. This is in accordance with recent molecular genetics evi-
ref. 49). Also, we expect that BMI itself and the neurobehavioral
dence that 75% of obesity-related genes express preferentially in
mechanisms behind it are continuum processes, therefore all vari-
the brain (4). Similarly, the genetic correlation between cogni-
ation in the range from normal weight to obesity is likely helping to
tion and BMI uncovered in our sample is at the same magni-
tude as molecular estimates of associations between more uncover underlying associations. While the measurement of cog-
specific cognitive measures and BMI (15, 18). The current evi- nition and personality was exhaustive, it lacked some common
dence further supports the brain–gene association with obesity behavioral tasks like the stop-signal task, or common questionnaires
vulnerability. measuring self-control, impulsivity, and eating-specific behaviors,
A possible explanation of the genetic correlations is pleiotropy— that have been previously associated with body weight (5, 6). Par-
the existence of a common set of genes that influence variance in ticularly, the common eating-specific behaviors such as un-
both obesity and brain function. It is possible that people with a controlled eating (50) are likely better candidates for explaining
higher genetic risk for obesity also have genetic propensity for the brain morphology–BMI associations as they are more directly re-
brain and cognitive patterns outlined here. It is also likely that in- lated to the hypothesized underlying behavior.
terventions could influence both obesity and brain function. For One has to be careful in translating individual differences in cor-
instance, regular exercise can support weight management (39), tical thickness in normal populations to underlying neural mecha-
reduce the heritability of obesity (40), and improve cognitive nisms. Diverse biological processes have been suggested to influence
health (41). MRI-based cortical thickness measures, ranging from synaptic den-
However, our results could also support a causal relationship— sity to apparent thinning due to synaptic pruning and myelination
that the genetic correlation is due to a persistent effect of heri- (summarized in refs. 51 and 52). A definitive model of the underlying
table brain factors on overeating and, hence, BMI. For instance, mechanism that links normal variations in cortical thickness to dif-
we could hypothesize that the heritable obesity-related cognitive ferences in brain function cannot be given, as cortical thickness has
profile promotes overeating when high-calorie food is available. As not been mapped with both MRI and histology in humans (52).
high-calorie food is abundant and inexpensive, the cognitive risk Still, the associations between cortical thickness and BMI in one
profile could lead to repeated overeating, providing an opportunity sample were able to predict BMI in a new separate sample, sug-
for genetic obesity proneness to express. Such longitudinal gesting that the pattern is robust. Our conceptual interpretation of
environmental effects of a trait need not to be large, they just the significance of cortical thickness patterns has support from
have to be consistent (ref. 42, see discussion in ref. 43). Of measures of both brain structure and cognitive function.
course, a reverse scenario is also possible—obesity leads to al- Relying on PPS-s prevented us from analyzing detailed inter-
terations in cortical morphology due to the consequences of actions between cortical thickness and cognitive function and
cardiometabolic complications, including low-grade chronic their genetic overlap with each other. However, given the rela-
inflammation, hypertension, and vascular disease (reviewed in tively small associations between PPS-s and the number of can-
refs. 9 and 44). However, we find this hypothesis less plausible didate measures that could be expected to interact with one
as global brain atrophy due to metabolic syndrome is mostly seen in another, we believe it would have been hard to find an associa-
older participants, whereas the current sample had a mean age of tion that would have survived multiple testing corrections.

Vainik et al. PNAS Latest Articles | 5 of 6

Future, focused, hypothesis-driven studies have to further elu- demographics, and family structure are summarized in SI Appendix, SI Methods
cidate the neurobehavioral mechanisms behind obesity proneness. and Table S1. Software pipelines for obtaining features of cortical thickness and
In summary, the current analysis provides comprehensive evi- brain volume are described in SI Appendix, SI Methods. Analysis scripts to re-
produce results presented are available at: https://osf.io/htx7u.
dence that the obesity-related differences in brain structure and
SI Appendix, Fig. S1 provides a schematic pipeline for data analysis. De-
cognitive tests are largely due to shared genetic factors. Genetic
tails of each data analysis step are outlined in SI Appendix, SI Methods. We
factors also explain occasional overlap between neurobehavioral describe how PPS weights are obtained through cross-validation and how
factors. We hope that increasingly larger longitudinal datasets and the weights generalize to a separate dataset (S1200n). We further describe
dedicated studies will help to outline more specific neurobehavioral the main principles of twin and sibling-based heritability analysis and rep-
mechanisms that confer vulnerability to obesity and provide a basis lication of these findings using individual features instead of PPS-s and
for designing informed interventions. replication in a separate dataset (S1200n). Finally, the software and pack-
ages used are listed.
Methods
Data were provided by the Human Connectome Project (24). Certain people were ACKNOWLEDGMENTS. Supported by funding from Canadian Institutes of
excluded due to missing data or not fulfilling typical criteria. Exclusion details, Health Research (A.D.) and Fonds de Recherche du Québec – Santé (U.V.).

1. Di Angelantonio E, et al.; Global BMI Mortality Collaboration (2016) Body-mass index 27. Mõttus R, Realo A, Vainik U, Allik J, Esko T (2017) Educational attainment and per-
and all-cause mortality: Individual-participant-data meta-analysis of 239 prospective sonality are genetically intertwined. Psychol Sci 28:1631–1639.
studies in four continents. Lancet 388:776–786. 28. Noble KG, et al. (2015) Family income, parental education and brain structure in
2. Hammond RA, Levine R (2010) The economic impact of obesity in the United States. children and adolescents. Nat Neurosci 18:773–778.
Diabetes Metab Syndr Obes 3:285–295. 29. Ad-Dab’bagh Y, et al. (2006) The CIVET image-processing environment: a fully au-
3. Elks CE, et al. (2012) Variability in the heritability of body mass index: A systematic tomated comprehensive pipeline for anatomical neuroimaging research. Proceedings
review and meta-regression. Front Endocrinol (Lausanne) 3:29. of the 12th Annual Meeting of the Organization for Human Brain Mapping (Elsevier,
4. Locke AE, et al.; LifeLines Cohort Study; ADIPOGen Consortium; AGEN-BMI Working New York), p 2266.
Group; CARDIOGRAMplusC4D Consortium; CKDGen Consortium; GLGC; ICBP; MAGIC 30. Collins DL, Evans AC (1997) Animal: Validation and applications of nonlinear
Investigators; MuTHER Consortium; MIGen Consortium; PAGE Consortium; ReproGen registration-based segmentation. Int J Pattern Recognit Artif Intell 11:1271–1294.
Consortium; GENIE Consortium; International Endogene Consortium (2015) Genetic 31. Val-Laillet D, et al. (2015) Neuroimaging and neuromodulation approaches to study
studies of body mass index yield new insights for obesity biology. Nature 518:197–206. eating behavior and prevent and treat eating disorders and obesity. Neuroimage Clin
5. Michaud A, Vainik U, Garcia-Garcia I, Dagher A (2017) Overlapping neural endo- 8:1–31.
phenotypes in addiction and obesity. Front Endocrinol (Lausanne) 8:127. 32. Corbetta M, Shulman GL (2002) Control of goal-directed and stimulus-driven atten-
6. Vainik U, Dagher A, Dubé L, Fellows LK (2013) Neurobehavioural correlates of body tion in the brain. Nat Rev Neurosci 3:201–215.
mass index and eating behaviours in adults: A systematic review. Neurosci Biobehav 33. Aminoff EM, Kveraga K, Bar M (2013) The role of the parahippocampal cortex in
Rev 37:279–299. cognition. Trends Cogn Sci 17:379–390.
7. Emery RL, Levine MD (2017) Questionnaire and behavioral task measures of impul- 34. Hargrave SL, Jones S, Davidson TL (2016) The outward spiral: A vicious cycle model of
sivity are differentially associated with body mass index: A comprehensive meta- obesity and cognitive dysfunction. Curr Opin Behav Sci 9:40–46.
analysis. Psychol Bull 143:868–902. 35. Neseliler S, Han J-E, Dagher A (2017) The use of functional magnetic resonance im-
8. Bartholdy S, Dalton B, O’Daly OG, Campbell IC, Schmidt U (2016) A systematic review aging in the study of appetite and obesity. Appetite and Food Intake: Central Control,
ed Harris RBS (CRC/Taylor & Francis, Boca Raton, FL), 2nd Ed.
of the relationship between eating, weight and inhibitory control using the stop
36. Jansen A, Houben K, Roefs A (2015) A cognitive profile of obesity and its translation
signal task. Neurosci Biobehav Rev 64:35–62.
into new interventions. Front Psychol 6:1807.
9. Willette AA, Kapogiannis D (2015) Does the brain shrink as the waist expands?
37. Doucet GE, Rasgon N, McEwen BS, Micali N, Frangou S (2017) Elevated body mass
Ageing Res Rev 20:86–97.
index is associated with increased integration and reduced cohesion of sensory-driven
10. Alonso-Alonso M, Pascual-Leone A (2007) The right brain hypothesis for obesity.
and internally guided resting-state functional brain networks. Cereb Cortex 28:
JAMA 297:1819–1822.
988–997.
11. Winkler AM, et al. (2010) Cortical thickness or grey matter volume? The importance
38. Vainik U, Mõttus R, Allik J, Esko T, Realo A (2015) Are trait–outcome associations
of selecting the phenotype for imaging genetics studies. Neuroimage 53:1135–1146.
caused by scales or particular items? Example analysis of personality facets and BMI.
12. Veit R, et al. (2014) Reduced cortical thickness associated with visceral fat and BMI.
Eur J Pers 29:688–634.
Neuroimage Clin 6:307–311.
39. Wadden TA, Webb VL, Moran CH, Bailer BA (2012) Lifestyle modification for obesity:
13. Medic N, et al. (2016) Increased body mass index is associated with specific regional
New developments in diet, physical activity, and behavior therapy. Circulation 125:
alterations in brain structure. Int J Obes 40:1177–1182.
1157–1170.
14. Mole TB, Mak E, Chien Y, Voon V (2016) Dissociated accumbens and hippocampal structural
40. Horn EE, Turkheimer E, Strachan E, Duncan GE (2015) Behavioral and environmental
abnormalities across obesity and alcohol dependence. Int J Neuropsychopharmacol 19:
modification of the genetic influence on body mass index: A twin study. Behav Genet
pyw039. 45:409–426.
15. Marioni RE, et al.; CHARGE Cognitive Working Group (2016) Assessing the genetic
41. Hillman CH, Erickson KI, Kramer AF (2008) Be smart, exercise your heart: Exercise
overlap between BMI and cognitive function. Mol Psychiatry 21:1477–1482. effects on brain and cognition. Nat Rev Neurosci 9:58–65.
16. Spieker EA, et al. (2015) Shared genetic variance between obesity and white matter 42. Dickens WT, Flynn JR (2001) Heritability estimates versus large environmental effects:
integrity in Mexican Americans. Front Genet 6:26. The IQ paradox resolved. Psychol Rev 108:346–369.
17. Rapuano KM, et al. (2017) Genetic risk for obesity predicts nucleus accumbens size 43. Tucker-Drob EM, Harden KP (2012) Intellectual interest mediates gene-by-SES in-
and responsivity to real-world food cues. Proc Natl Acad Sci USA 114:160–165. teraction on adolescent academic achievement. Child Dev 83:743–757.
18. Sanchez-Roige S, et al.; 23andMe Research Team (2018) Genome-wide association 44. Guillemot-Legris O, Muccioli GG (2017) Obesity-induced neuroinflammation: Beyond
study of delay discounting in 23,217 adult research participants of European ancestry. the hypothalamus. Trends Neurosci 40:237–253.
Nat Neurosci 21:16–18. 45. Cheng HL, et al. (2013) Iron, hepcidin and inflammatory status of young healthy
19. Lancaster TM, Ihssen I, Brindley LM, Linden DE (2018) Preliminary evidence for genetic overweight and obese women in Australia. PLoS One 8:e68675.
overlap between body mass index and striatal reward response. Transl Psychiatry 8:19. 46. Cheng HL, Medlow S, Steinbeck K (2016) The health consequences of obesity in young
20. Opel N, et al. (2017) Prefrontal gray matter volume mediates genetic risks for obesity. adulthood. Curr Obes Rep 5:30–37.
Mol Psychiatry 22:703–710. 47. Engvig A, et al. (2010) Effects of memory training on cortical thickness in the elderly.
21. Weise CM, et al. (2017) The obese brain as a heritable phenotype: A combined Neuroimage 52:1667–1676.
morphometry and twin study. Int J Obes 41:458–466. 48. Appelhans BM, French SA, Pagoto SL, Sherwood NE (2016) Managing temptation in obesity
22. Aron AR, Robbins TW, Poldrack RA (2014) Inhibition and the right inferior frontal treatment: A neurobehavioral model of intervention strategies. Appetite 96:268–279.
cortex: One decade on. Trends Cogn Sci 18:177–185. 49. CDC (2017) New adult obesity maps. Cent Dis Control Prev. Available at https://www.
23. Allen TA, DeYoung CG (2017) Personality neuroscience and the five factor model. The cdc.gov/obesity/data/prevalence-maps.html. Accessed March 26, 2018.
Oxford Handbook of the Five Factor Model (Oxford Univ Press, Oxford). 50. Vainik U, Neseliler S, Konstabel K, Fellows LK, Dagher A (2015) Eating traits question-
24. Van Essen DC, et al.; WU-Minn HCP Consortium (2013) The WU-Minn human con- naires as a continuum of a single concept. Uncontrolled eating. Appetite 90:229–239.
nectome project: An overview. Neuroimage 80:62–79. 51. Fjell AM, et al. (2015) Development and aging of cortical thickness correspond to
25. Kim J (2016) Personality traits and body weight: Evidence using sibling comparisons. genetic organization patterns. Proc Natl Acad Sci USA 112:15462–15467.
Soc Sci Med 163:54–62. 52. Walhovd KB, Fjell AM, Giedd J, Dale AM, Brown TT (2017) Through thick and thin: A
26. Ritchie SJ, Tucker-Drob EM (2018) How much does education improve intelligence? A need to reconcile contradictory results on trajectories in human cortical development.
meta-analysis. Psychol Sci 29:1358–1369. Cereb Cortex 27:1472–1481.

6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.1718206115 Vainik et al.